Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "The shape of galaxy disks: how the scale height increases with galactocentric distance"

See other formats

A&A manuscript no. 

(will be inserted by hand later) 

Your thesaurus codes are: 

1(11.06.2; 11.16.1; 11.19.2; 11.19.6; 11.19.7) 





Letter to the Editor 

The shape of galaxy disks: how the scale height increases 

with galactocentric distance* 

R. de Grijs and R.F. Peletier 

Kapteyn Astronomical Institute, University of Groningen, P.O. Box 800, 9700 AV Groningen, The Netherlands 

^^ ' Received: November 5, 1996; accepted: February 21, 1997 












Abstract. We present the results of a detailed study of 
vertical surface brightness profiles of edge-on disk galax- 
ies. Although the exponential disk scale height is constant 
to first order approximation, we show that for the large 
majority of galaxies in our sample, the scale height in- 
creases with distance along the major axis. The effect is 
strongest for early-type galaxies, where the increase of the 
scale height can be as much as a factor of 1.5 per scale- 
length, but is almost for the latest-type galaxies. The 
effect can be understood if early-type disk galaxies have 
thick disks with both scale lengths and scale heights larger 
than those of the dominant disk component. Its origin ap- 
pears to be linked to the processes that have formed the 
thick disk. 

Key words: galaxies: fundamental parameters — galax- 
ies: photometry — galaxies: spiral — galaxies: statistics 
— galaxies: structure 

1. Disk vertical scale parameters 

The behaviour of the thickness of galactic disks is of ma- 
jor importance for understanding the evolution of disk 
galaxies. Since the exponential vertical scale height can 
be related directly to the vertical velocity dispersion in 
(isothermal) disks (e.g., Bahcall & Casertano, 1984), we 
may be able to constrain models that describe the dynam- 
ical heating mechanisms of galactic disks by studying this 

Van der Kruit & Searle (1981a,b; 1982) found, for their 
sample of edge-on spirals, that the vertical scale param- 
eter, zq = 2 hz, is in good approximation independent 
of position along the major axis. Later studies of NGC 

Send offprint requests to: R. de Grijs 

* Based on observations obtained at the European Southern 

Observatory, La Silla, Chile 

891 (Kylafis & BahcaU, 1987) and NGC 5907 (Barnaby 
& Thronson, Jr., 1992) have confirmed this result. From 
two-dimensional modeling of a sample of 10 edge-on spiral 
and lenticular galaxies, Shaw & Gilmore (1990) found that 
the radial variation of scale heights is typically within ± 
3% of the derived mean for the main disk component, with 
no obvious dependence on colour or model type adopted. 

1.1. Why should the scale height be constant? 

Van der Kruit & Searle (1982), following a suggestion by 
Fall, pointed out that the scale height would be constant 
during the secular evolution of a galactic disk if: 

— the disk is continuously heated by, e.g., the random 
acceleration of the disk stars by giant molecular clouds 
or spiral structure (Spitzer & Schwarzschild, 1951). 

— at all times the star formation rate is proportional 
to the surface number density of the giant molecular 

In this case the radial distribution of both the verti- 
cal velocity dispersion and the surface mass density of the 
luminous matter would be determined by the radial distri- 
bution of the giant molecular clouds. Since the vertical ve- 
locity dispersion and surface mass density possess similar 
radial distributions, their ratio has no radial dependence. 
The scale height also will not exhibit a dependence on ra- 
dial distance, since it is proportional to this ratio (van der 
Kruit & Searle, 1981a; van der Kruit, 1988). 

However, at present there is no satisfactory expla- 
nation as to how disk heating, the rate of which must 
vary greatly with radius from the observed distribution 
of molecular clouds in our own and other galaxies, can 
naturally lead to this result. 

Alternatively, the so-called "self regulation" mecha- 
nism for galaxy disks was used to explain the constancy of 
the vertical scale parameter (see Bertin & Lin, 1987). This 
mechanism, which describes the disk evolution, leads to a 
more or less constant value for Toomre's (1964) Q param- 

R. de Grijs and R.F. Peletier: The shape of galaxy disks 

eter as a function of radius. Combined with an exponen- 
tially decreasing velocity dispersion, as found in exponen- 
tial disks, this leads to a constant scale height (Bottema, 

However, on closer inspection the constancy of the 
scale height seems to lose strength in the (radially) outer 
parts in a number of galaxies studied. For our Galaxy, 
Kent et al. (1991) find indications that the scale height 
increases linearly with radius from a minimum radius out- 
wards. The improvement over models with a fixed scale 
height is significant, although not dramatic. 

1.2. A statistically complete sample 

In order to be able to draw conclusions based on statistical 
considerations rather than on individual cases we need a 
large, homogeneous data set, selected in such a way that 
unwanted selection biases are avoided. 

A diameter-limited sample, taken from the Surface 
Photometry Catalogue of the ESO-Uppsala Galaxies 
(ESO-LV, Lauberts & Valentijn 1989) was used. 

We selected all non-interacting galaxies for which ei- 
ther the blue angular diameter (at a surface brightness 
level of hb = 25 mag arcsec"^), Df^, as measured by an 
automatic ellipse fitting routine, or the original visual di- 
ameter, Dyis, is greater than 2. '2. From V/Vmax complete- 
ness tests (e.g., Davies, 1990) it follows that the ESO-LV 
is statistically complete down to blue angular diameters, 
D^ and Dyis combined, of l.'O. 

To the diameter-limited sample obtained we applied an 
inclination selection criterion. As van der Kruit & Searle 
(1981a) showed, for systems having inclinations i > 86° 
the slopes of both the radial and the vertical structures are 
independent of inclination. Therefore we defined our lower 
inclination limit to be 87°. Inclinations were determined 
following de Grijs & van der Kruit (1996). 

For the observed sample the mean V/Vmax value re- 
sults in 0.473 ± 0.041, so that we conclude that it is at 
least statistically complete. 

From the total sample of 93 edge-on galaxies we ran- 
domly selected 48, which were observed in optical pass- 
bands {B, V, R, and /) using both the Danish 1.54m and 
the Dutch 0.92m telescopes of the European Southern Ob- 
servatory at La Silla, Chile, from 1993 to 1996. Both tele- 
scopes were used in direct imaging mode, at prime focus. 

2. Results 

We extracted vertical luminosity profiles at a number of 
positions along the major axes of the sample galaxies. 
To do this, the galaxies were binned semi-logarithmically 
both radially and vertically to retain an approximately 
constant overall S/N ratio in the resulting vertical profiles 
and to be able to follow the profiles further out. 

In the analysis of the individual profiles, we distin- 
guished between the different sides of the galaxy planes. 

to avoid possible dust contamination in the case of not 
perfectly edge-on orientations. 

Both van der Kruit (1988) and de Grijs & van der 
Kruit (1996), have shown that the vertical profiles are 
more peaked than expected from an isothermal distribu- 
tion, but less peaked than exponential, probably close to 
a sech(2;) distribution. The analysis shows that we can 
approximate the profile's shape with an exponential dis- 
tribution at 0-heights greater than 1.5 scale heights. We fit 
the vertical profiles out to 4 scale heights, thereby taking 
into account the possible presence of underlying thick disk 
components. Comparison for 24 galaxies with our unpub- 
lished K'-h&nd images showed that in this vertical region 
the contamination of the stellar light by dust extinction 
can also be considered to be negligible. 

_ 1 1 II 1 1 1 1 1 1 1 1 1 

'i\ ' '-_ 

:' 'J 

1 1 1 1 1 1 1 1 1 

1 ' '; 



i ^ 

T- CD 

1 i 

■ — 


? 10 

^ • * 





il 5 





:x^,^ •, , , , , 

1 , ,- 



-4-2 2 

4 -4 

-2 2 



1 1 1 1 1 1 1 1 1 1 1 1 1 1 



1 1 1 1 1 1 1 1 1 1 

1 ' '- 

S 8 









^J; 6 

*. i. 





X 4 





7J ^ 

'-^"^ 1 ■ 

III 1 1 1 1 1 1 1 1 1 1 1 1 



1 1 1 1 1 1 1 1 1 

1 , ,- 


-4-2 2 

4 -4 

-2 2 



_ 1 1 

1 1 1 1 1 1 1 1 1 1 1 

1 1 1 _ 

E' ' 1 

1 1 1 1 1 1 1 1 1 

1 1 1: 


5 ^ 



5 c 



X 4 



^ . s 8 






- ■ a sHo* * ' 

(D _" 


'-(""^ 1 1 

1 1 1 1 1 1 1 1 1 1 1 1 1 



1 1 1 1 1 1 1 1 1 

1 1 i: 


-4-2024 -4-2024 

Posil.ion along the major axis (R / hp) 

Fig. 1. 

Examples of the /-band scale height behaviour as a function of 
galactocentric distance for (a) ESQ 358G-29 (T = -1.6), (b) 
ESQ 311G-12 (T = 0.0), (c) ESQ 315G-20 (T = 1.0), (d) ESQ 
322G-87 (T = 3.4), (e) ESQ 435G-50 (T = 5.0), and (f) ESQ 
505G-03 (T — 7.7). Open and closed symbols represent data 
taken on both sides of the galaxy planes. 

We note that the scale height seems to increase with 
radius along the major axis, depending on galaxy type. As 
an example, in Fig. 1 we show representative scale height 
distributions as a function of position along the galaxies' 
major axes for galaxies representing various types. 

We fit the scale height distribution as a function of 
position along the major axis for the total range of data 
points, between 1.5 and 3.5, and between 2.0 and 4.0 I- 
band scale lengths. The results obtained for the fitting 
range between 2.0 and 4.0 scale lengths do not differ sub- 
stantially from the fitting range between 1.5 and 3.5 scale 

„ 0.2 ^ — |- 
cc 1(a) 

\ 0-1 r--*i--i-__ 

R. de Grijs and R.F. Peletier: 

iH-i .i,i 

J^ '-y= -0.0106 ( + 0.0007) '1' + 0.0745 (-0.0036) 


I band - 



f 1 ~ ~'i ■^*-- A*-^._- 


17= -0.00H6 ( I 0.0012) T I 0.0711 (-0.0055) 


2 4 

Revised Hubble Type (T) 

Fig. 2. 

Disk scale height gradients as a function of revised Hubble 
type. The closed symbols represent the data; the open symbols 
are type-averaged data points used to determine the relation- 
ship described in the text. The results obtained for NGC 891 
are shown by crosses (x). For comparison, the best fit obtained 
for the B-band data is also shown in the /-band figure 

The shape of galaxy disks 

ESQ 358G-26 (T -2.0) 

16 ^ ; 

IB |1 • 

\ \ 


26 ^^ 



.— 18 

f. 20 

& 22^ 


ESQ 500G-24 (T = -1.0) 

'\' ' r ' 1 ' I V ' '\' h' ' ' ' I ' '^ 

\ \ \ \ \ \ (cj - 

; : \ : \ \ 

• • • '. V \ V 

c r. •. r. \ \V 

' ? * > > v • 

10 20 30 40 

ESQ 321G 10 (T = 0.0) 
1 B b , .. \ \ (e) - 




18 4 

22 \\ ^ \\ 

ESQ 358C-29 (T = -1.6) 

(aj^ 16 

20 rt • '4 1 1 ' > H 20 
22 -I ; • \ » \ : ^ 22 

•^ 26 t 


\ \ \ 

\ \ 





ESQ 311G-12 (T = 0.0) 

'l ,\. .\. l\. ). . .^ 
10 20 30 

ESQ 435G-50 (T = 5.0) 
.|- 4 \ \ \ \ ('/ — 

\ A \ 

\ \ : t \ ^ 

N. • ^. V * » 

z height (h^^) 

lengths, whereas the scatter increases if we take all data 
points, including those dominated by the bulge light con- 

Fig. 2 shows the results for both the B and the I- 
band data, for the fitting range between 1.5 and 3.5 scale 
lengths. We see that they do not differ significantly. This 
means that the effect is intrinsic to the old disk, and not 
affected by, e.g., dust or a young stellar population. Es- 
pecially for the earlier-type galaxies of our sample the ob- 
served relationship might be due to a non-negligible bulge 
contribution. To discard this possibility, we varied our ra- 
dial fitting range. We estimated the bulge influence by 
applying a two-dimensional bulge-disk decomposition to 
the data. For this sample, the bulge contribution can be 
assumed to be negligible (i.e. less than 5% of the total 
hght) outside 1.8 ± 0.3 /-band scale lengths, depending 
on galaxy type. 

To test our method, we applied the same procedure to 
the data of van der Kruit & Searle (1981b) for the edge-on 
galaxy NGC 891, and found that the slopes obtained for 
NGC 891 fall in the same range as our more recent data. 
NGC 891 is also included in Fig. 2. 

To understand better the nature of this effect we show 
for a few representative galaxies vertical profiles at various 
radial distances from the center in Fig. 3. In the case of 
ESO 358G-29 a bright thick disk is seen, with scale length 
larger than the thin disk. This fact that the scale lengths 
are different causes the apparent increase of the galaxy's 
scale height with radius. In the other early-type spirals, 
except 1, the thick disk cannot be seen so clearly, although 
the scale height here increases with radius. We argue that 
the increase of scale height with radius is linked to the 
presence of a thick disk, and with galaxy type, since both 
effects only occur in early-type spirals. Based on Fig. 2 it 
could be that all spirals of type earlier than 2 have a thick 

Fig. 3. 

(a) - (f) Examples of increasing scale height with galactocen- 
tric distance, compared to fixed scale height models. The fit- 
ting of the slopes was done between 1.5 and 4 scale heights. 
From left to right the profiles that are shown were extracted 
at logarithmically increasing galactocentric distances 

disk with scale length larger than the thin disk. To check 
this we have decomposed all galaxies with type < 0.0 into 
2 components with constant scale height. We obtain good 
fits for all galaxies except for ESO 321G-10, with scale 
height ratios ranging from 1.8 to 4.6. 

3. Discussion 

Previous studies have always resulted in confirmations of 
the assumed constancy of the vertical scale height, al- 
though in some cases authors refer to the fact that the 
outermost profiles in their sample galaxies indicate smaller 
slopes and hence larger scale heights, e.g. van der Kruit 
& Searle (1981a) for the late-type edge-on galaxies NGC 
4244 and NGC 5907 and de Grijs & van der Kruit (1996) 
for a number of their earlier- type sample galaxies. 

Barnaby & Thronson, Jr. (1992) find for NGC 5907 
that the scale height is lower between —100" and -1-100" 
than in the rest of the galaxy. They say that this is due to 
bulge contamination. However, as can be seen from their 
Fig. 4, between ±50" and ±100" the bulge contribution is 

Van der Kruit & Searle (1981a) explained these obser- 
vations by invoking a possible thickening of the disk just 
before the radial cut-off they observed or by attributing 
it to the influence of an optical warp. However, even if 
the shallowing at the last profiles were entirely due to a 

R. de Grijs and R.F. Peletier: The shape of galaxy disks 

thickening of the disk the imphed change in zq would be 
relatively small and restricted to no more than 10% of the 
observable extent of the galaxies. It should be noted that 
in a number of profiles their model deviates significantly 
from the data, although they calculate a single scale height 
and fit all profiles with a constant vertical scale parameter. 

In general, thick disks seem to occur more often in 
early-type galaxies than in later types (e.g., Burstein, 
1979). We have found that the exponential vertical scale 
height increases as function of position along the major 
axis and that this dependence is strongest for the earliest- 
type galaxies. It seems therefore a natural theory to link 
this increase of scale height with radius to the presence of 
a thick disk. 

For that reason, it is likely that the formation mech- 
anism for the thick disk and the origin of the increas- 
ing scale height effect are similar. The possibility that the 
thick disk is the intermediate component in the hierarchi- 
cal formation scenario between the bulge and the thin disk 
(as suggested by, e.g., Gilmore, 1984) seems to be ruled 
out by the large scale length of the thick disk. Therefore, 
more likely formation mechanisms for the thick disk invoke 
the accretion of material by the early thin disk, causing 
violent dynamical heating processes to take place, thereby 
pufHng up the thin disk (e.g., Norris, 1987; Statler, 1989). 

Quinn et al. (1993) find that disks following accretions 
are noticeably fiared. The projected and deprojected scale 
heights of both the original disk and of the final system 
increase with radius. Quinn et al. (1993) have also shown 
that the disk is heated in all three directions by the sim- 
ulated merger event and that in the outer regions the ra- 
dial velocity dispersion is nearly constant, while the verti- 
cal dispersion rises. Van der Kruit & Freeman (1986) and 
Bottema (1993) have studied a number of disk galaxies, 
especially of later type, in detail, and found that the veloc- 
ity dispersions as a function of radius are consistent with 
a constant Q parameter (Toomre, 1964), i.e. their mea- 
surements are consistent with an exponentially decreasing 
vertical velocity dispersion, and assuming a fiat rotation 

Bertola et al. (1995) and Fisher (1997) find for their 
samples of early-type disk galaxies that the stellar veloc- 
ity dispersions may decrease less rapidly than exponen- 
tially, along both the major and minor axes. This result 
is independent of viewing angle and would indicate that 
the radial and vertical velocity dispersions become more 
nearly constant with radius. 

These results could imply that our observed increas- 
ing scale height with galactocentric distance may be re- 
lated to the stellar velocity dispersion falling more slowly 
than predicted by the constant Q model. Similarly, a non- 
exponential behaviour of the stellar velocity dispersion is 
suggestive of a non-constant disk scale height, or an in- 
creasing thick disk contribution to the mass profile. 

4. Conclusions 

— We have found that, although constant in a first order 
approximation the exponential vertical scale height in- 
creases as a function of position along the major axis 
and this effect is strongest for early-type disk galaxies, 
where this increase can be as much as a factor of 1.5 
per scale length. 

— This effect might be closely related to the presence 
of thick disks, with a scale length larger than that of 
the thin disk. The fact that, in this case, the scale 
length of the thick disk has to be larger than that of 
the thin disk strengthens the possibility that the thick 
disks were made by merging of small satellites (see, 
e.g., Quinn et al., 1993). 

— If, as theory predicts, the scale height is closely coupled 
to the vertical velocity dispersion, one should see a less 
rapid decline of the vertical velocity dispersion than an 
exponential decline. 

Acknowledgements - We thank Piet van der Kruit, Roelof 
Bottema, David Fisher, Linda Sparke and the referee 
for the useful discussions and suggestions concerning this 


Bahcall, J.N., Casertano, S., 1984, ApJ 284, L35 

Barnaby, D., Thronson Jr., H.A., 1992, AJ 103, 41 

Bertin, G., Lin, C.C, 1987, in: Proceedings of the 10th lAU 

European Regional Astronomy Meeting, Prague, p. 255 
Bertola, F., Cinzano, P., Corsini, E.M., Rix, H.-W., Zeilinger, 

W.W., 1995, ApJ 448, LIS 
Bottema, R., 1993, A&A 275, 16 
Burstein, D., 1979, ApJ 234, 829 
Davies, J.L, 1990, MNRAS 244, 8 

de Grijs, R., van der Kruit, P.C., 1996, A&AS 117, 19 
Fisher, D., 1997, AJ, in press 
Gilmore, G., 1984, MNRAS 207, 223 
Kent, S.M., Dame, T.M., Fazio, G., 1991, ApJ 378, 131 
Kylafis, N.D., Bahcall, J.N., 1987, ApJ 317, 637 
Lauberts, A., Valentijn, E.A., 1989, The Surface Photometry 

Catalogue of the ESO-Uppsala Galaxies (ESO-LV), ESQ 
Norris, J., 1987, in: The Galaxy, eds. Gilmore &: Carswell, p. 

Quinn, P.J., Hernquist, L., FuUager, D.P., 1993, ApJ 403, 74 
Schweizer, F., Seitzer, P., 1992, AJ 104, 1039 
Shaw, M.A., Gilmore, G., 1990, MNRAS 242, 59 
Spitzer, L., Schwarzschild, M., 1951, ApJ 114, 385 
Statler, T.S., 1989, ApJ 344, 217 
Toomre, A., 1964, ApJ 139, 1217 
van der Kruit, P.C., 1988, A&A 192, 117 
van der Kruit, P.C., Freeman, K.C., 1986, ApJ 303, 556 
van der Kruit, P.C., Searle, L., 1981a, A&A 95, 105 
van der Kruit, P.C., Searle, L., 1981b, A&A 95, 116 
van der Kruit, P.C., Searle, L., 1982, A&A 110, 61 

This article was processed by the author using Springer- Verlag 
ET^ A&A style file L-AA version 3.